Pruning the decision tree

Fighting is a messy endeavor, and one could argue the whole point of martial arts has always been to rationalize it in a way that could be explained, taught, and reproduced. To this end, guards, blows and plays have been designed to describe in a succinct manner a possibly infinite collection of fencing exchanges. A system then allows a fencer to assess the situation they’re in and make an informed decision based on the relevant stimuli they’ve interpreted.

This tactical process is often modeled as a decision tree: a fencing exchange consists in a series of logical decisions based on the opponent’s reaction. The goal of a HEMA instructor according this pedagogical framework is therefore to teach their students the proper pattern recognition skills as well as the adequate counters. However, does this model actually match the fencing art taught by Fiore de’i Liberi?

Simple situation, complicated answers

Let’s ignore Fiore’s armizare for a moment and, as a thought experiment, try to figure out what would be an intuitive follow-up to this common sparring situation: starting from your dominant side, you have beaten the opponent’s blade away. This beat may have been a parry against a cut, or a preparatory action against an extended sword; we will focus on the latter case for simplicity’s sake, but it does not make much of a difference, if any. It turns out clearing the opponent’s sword leaves an obvious open path for a direct false edge mezzano to the head with a passing step.

A common reaction to such a sharp pressure in an actual sparring situation is to lift the hands and almost fortuitously parry the incoming false edge cut.

As a consequence, the fencer initiating the beat has to look for another opening – they may use, as an example, an indirect rising cut to the arms, the elbow, or the armpit.

However, a more experienced fencer can yield to the beat and use the momentum to chamber a blow such as a fendente to the head. The likely result of this counter-offensive action is a double hit.

To counter this cut around, the beat has to be followed immediately by a parry (performed here with the false edge), then a riposte once the opponent’s incoming attack has been properly dealt with. Note that the passing step has to be aborted in order to avoid walking right into the incoming attack.

What ought to be a simple situation is quickly devolving into a sprawling decision tree. Knowing how to answer every single move the opponent can perform with a tailored, perfect counter is of very little use if you have trouble determining which technique the opponent will try in the first place – something that is easier said than done, as the three situations outlined previously share the same tactile feedback.

Our technical framework fell prey to overspecialization: to each action, we paired an optimal counter, but we did not take into account the burden of split-second decision making based on partial information. As a consequence, a tiny slip in one’s assessment of the opponent’s behavior may lead to an utterly inappropriate reaction. Fortunately, it turns out our beloved fencing master had a practical answer to this crippling issue.

A simple play for simple minds

Fiore’s third play of zogho largo, the crossing at the middle of the swords, may indeed provide us with actionable advice that we can apply to this problematic situation.

[Getty, 25v-a] Here, we are also crossed in wide play, but at the middle of the sword. As soon as the cross is made, I let my sword glide over his hands; if I pass out of line with my right foot, I can push a thrust to the opponent’s chest, as you will see pictured right after this. (tr. by Tom Leoni and Greg Mele)

[Getty 25r-b] I have completed the play of my Master by performing the parry and immediately doing what he said: first, I wounded the opponent in the arms, then I placed a thrust into his chest. (tr. by Tom Leoni and Greg Mele)

The canonical context of this crossing isn’t really clear: Guy Windsor and Akademia Szermiezy favor a rising percussive parry from a low guard against an incoming cut; the Armizare society would rather use a counter-cut from posta di donna; I have dabbled with preparatory beats against an extended blade. In my experience, the exact setup doesn’t make much of a difference, as long as you are free to leave the bind safely afterwards.

In any case, once the opponent’s blade has been cleared, you are free to perform the canonical play and footwork: aim for the opponent’s left arm with a small fendente-like motion, then thrust over their arm. At the moment, this technique may seem convoluted, given the wide opening available.

However, as discussed earlier, the opponent is unlikely to remain passive if their blade has been beaten away. But lifting their hands is of very little help here: the left arm remains a vulnerable target, and you don’t even have to alter your motion if you perform the technique properly, passing to the opponent’s left.

Finally, if the opponent yields to the beat and chambers a blow, striking their left arm with a tight fendente suppresses and stifles the incoming cut in mid-motion. This technique is performed here with the last third (the tutta spada) of a blunt sword in a manner similar to the slices favored by the Liechtenauerian tradition. However, it seems likely that performing the same motion with the first third of a sharp sword, as depicted by Fiore, would still result in a pinning effect and may even create an opening for a thrust.

A single technique can therefore handle a wide variety of situations, even if it isn’t necessarily the optimal counter in any of the use cases we explored and may even seem needlessly complicated when the opening is obvious.

One size fits all

It may indeed be argued that we are over-interpreting the tactical implications of a simple largo play. However, Fiore de’i Liberi does explicitly value the sheer versatility of some of his favored techniques. Among them, rising beats from low guards may be some of the most common armizare parries performed with a wide variety of weapons against all sorts of blows.

[Getty 31r-b] I’ll wait for my opponents one by one, without fear of failure from any cut, thrust, or hand-held weapon thrown at me. I’ll advance the front foot, pass at an angle against the opponent’s weapon and beat it to his left side. After making my parry, I’ll instantly attack. (tr. by Tom Leoni and Greg Mele)

All it takes to parry any incoming strike is a last minute correction mentioned in the horseback section.

[Getty 43v-a] Bear in mind that thrusts and riversi must be beaten to the outside, that is, sideways, and not upward; fendenti should similarly beaten to the outside, lifting slightly the opponent’s weapon.

[Getty 43v-b] Bear in mind that this guard counters all the blows both on the mandritto and the riverso side, and is usable against right- or left-handed opponents. (tr. by Tom Leoni and Greg Mele)

This sweeping, rising beat from the left can intercept any strike if it is timed well. As a consequence, it may be less vulnerable to feints than a simple block from a centered, point in line guard: while the latter is a tighter and simpler motion, it requires a fencer to predict at the right time the exact opening the opponent will be aiming for, unlike Fiore’s favorite parry that is only timing-dependent. In a similar fashion, modern olympic fencers often rely on powerful sweeping parries that move through multiple lines, using several sweeps in a row if needs be.

A good technique should therefore strive to cover as many use cases as is reasonably possible, even if it isn’t necessarily the optimal action at any given point: a reliable, versatile parry remains more valuable than a fickle single-time counter. By unloading a fencer’s cognitive load, a good system allows them to focus on a few core motions and their tactical setup so they can be executed without hesitation, at the right measure, with speed and strength.

One could even argue the same principle can be extended to Fiore’s entire fencing system: through the use of guards, a skilled armizare fencer can significantly reduce the amount of tactical decisions they have to make. But this is a topic for another article…

Looking for Fiore’s tactics

Reconstructing historical motions and plays is one thing. Being able to explain them using a coherent tactical and technical framework is another. It has often been claimed that Fiore’s writings are at best fairly light in terms of tactical concepts, at worst, a mere bag of tricks. Does this unfortunate interpretation hold up to deeper scrutiny?

To answer this question, we need to have a look at one of the core theoretical concepts of Fiore’s art: the postures, also known as guards. A common interpretation is to treat guards solely as the starting and ending point of blows. Many later XVIth century authors, be it Joachim Meyer or Giovanni dall’Agocchie, often describe complex fencing actions as series of strikes that move through various positions known as guards. However, can the same be said of Fiore? The definition he introduces in the prologue isn’t very helpful:

[Getty 3r] A guard, or posta, is what you use to defend or guard yourself against the opponent’s attack. A posta, or guard, is a posture against the opponent, which you use to injure him without danger to yourself. (tr. by Tom Leoni and Greg Mele)

In order to form an opinion, we will take a deeper look at the actual description of  a few important postures.

The Full Iron Gate

We will first consider one of the main two-handed sword guards: tutta porta di ferro, the full iron gate.

[Getty 23v-a] The first one is Tutta Porta di Ferro (Full Iron Gate), very strong and good for waiting against any hand-held weapon (long or short), provided that your sword is of good quality and not too long. This guard parries, passes, and comes to the close. She can exchange thrusts and delivers her own. She can also beat thrusts to the ground, always proceeding with a pass and parrying any kind of attack. She can defend without much effort against anyone who picks a fight with her. (tr. by Tom Leoni and Greg Mele)

There is a a lot to unpack there. However, it should be fairly obvious now that this posture is much more than a position you move through during a sequence of cuts. Fiore actually introduces a fair amount of down to earth tactical advice we can reasonably read and interpret as follows:

  • Stand with your blade refused and wait for the opponent’s attack, then parry the committed blow you have provoked. You can also beat or engage an extended blade.
  • If you are outreached, establishing a firm bind will allow you to close in and get the upper hand.
  • As the opponent attacks, step into their strike to wrestle. Novice fencers too eager to wrestle often try instead to rush forward and attack a retreating opponent; this wild goose chase usually ends up with the attacker getting nailed by a nasty hand strike or stop hit. But the safest way to close in is often to step at the same time as the opponent: they won’t be able to retreat out of a move they actually initiated.

A quick summary would be: if your reach is shorter, you like to wrestle, and you don’t mind defending, then use the full iron gate. This posture supports a proactive, actionable tactical plan based on your own preferences.

The Boar’s Tooth

Let us introduce another of Fiore’s favourites: dente di zengiaro, the boar’s tooth. We may also study its close sibling, the middle boar’s tooth, that acts similarly in Fiore’s own words.

[Getty 24r-d] This is Dente di Cinghiaro (Boar’s Tooth), since it learned its offenses from the boar. It can deliver strong underhand thrusts all the way to the opponent’s face without passing; it then comes back down with a fendente to the arms. Sometimes, it can deliver a thrust to the opponent’s face, point up, while performing a quick accrescimento with the front foot, and recover back in guard with a fendente to the head and arms; then, it immediately delivers another thrust with the front foot accrescimento. It defends well against the close play. (tr. by Tom Leoni and Greg Mele with a correction of mine)

[Getty 24v-d] This is middle Dente di Cinghiaro. There are two Dente di Cinghiaro guards: the whole Dente di Cinghiaro and the middle Dente di Cinghiaro. This one is called “middle” because it is formed in the center of the body. What the other Dente di Cinghiaro can do, this can do. And just like the boar strikes at an angle, from this guard the sword always strikes at an angle to the opponent’s blade. From this guard, always attack with thrusts, uncover your opponent, mess up his hand and sometimes his head and arms. (tr. by Tom Leoni and Greg Mele)

How can we interpret this treasure trove of fencing actions?

  • Offensive actions from boar’s tooth should use lunging footwork instead of passing steps, trading raw power and explosive gap-closing ability for granular distance management as well as hit-and-run tactics.
  • Bind denial is important. Don’t let the opponent establish a firm crossing. Parry with strong beats and riposte immediately, void and hand snipe, throw a high thrust then hit with a cut, but avoid getting locked in a deep bind at all cost.
  • Stay on the move and aim for the opponent’s hands as often as you can. A fencer too eager to seek a narrow crossing will often expose their hands as they rush in and strike or thrust out of range. Punish this mistake.
  • The point of all these actions is to defend oneself from close play.

If you fear being overwhelmed by a stronger opponent as they close in, you may fight from the boar’s tooth and keep your distance using beats, feints, hand snipes and lunges. Not only do postures support our own fighting style, they may also be used to deny the opponent’s area of excellence.

The Long Position

Extended postures that threaten with the point are a staple of thrust fencing. Let’s take a look at Fiore’s posta longa, the long position.

[Getty 24r-a] This is the Posta Longa (Long Position), full of deception. She can probe the opponent’s guard to see if she can deceive him. If it is possible to strike with a thrust, she knows how to do it; she also know how to avoid cuts, then deliver them when it is possible. More than other guards, she can employ deception. (tr. by Tom Leoni and Greg Mele)

[Novati 18B-a] I am the Extended Stance with my short sword
And I often strike the throat with cunning.
(tr. by Michael Chidester)

Again, Fiore provides us with far more than a mere static posture. What can this guard do?

  • Bait the opponent by extending your blade. If you creep in close enough while threatening the opponent with the point, they’ll have to react sooner or later if they want to avoid being hit.
  • Probe your opponent’s defense before committing to a strike. Throwing a random cut and hoping that you end up in a good crossing you can work from isn’t a proper fencing strategy and will instead lead to sloppy exchanges. Good attacks have instead to be set up using preparatory actions in order to secure an opening.
  • Thrust if the opportunity arises. If your opponent does not react to your probing preparatory action, you should be able to promptly launch an actual attack.
  • Avoid blade contact if the opponent tries to beat your extended sword, then thrust or cut if there is enough time to do so. This is one of the most ubiquitous fencing techniques: disengage below or over the opponent’s blade, then strike again.
  • Use a shorter weapon. Shorter blades are more nimble and have an easier time disengaging than heavier, longer swords.

Fencing is first and foremost a dialogue between two fighters. It is far too common to obsess over technical details and forget the point of the whole art. Fiore emphasizes here the tactical qualities of the long posture: being able to gauge and deceive one’s opponent. He seldom wastes time on the technical minutiae of the motions performed.

Approaching your opponent with guards

You shouldn’t approach a fight without a plan based on an assessment of your own strengths and your opponent’s. Indeed, as early as the prologue to the wrestling chapter, Fiore warns the reader that:

[Getty 3r] When you engage in abrazare, you must assess whether your opponent is stronger or bigger than you, and whether he is much younger or older. You also need to take note of whether he places himself in any of the guards of abrazare. (tr. by Tom Leoni and Greg Mele)

As far as fencing with the two-handed sword is concerned, our plan will also depend on the following criteria:

  • Would you rather attack or defend?
  • Is your reach longer or shorter than your opponent’s?
  • Do you want to wrestle or would you rather keep your distance?
  • Would you rather feint or engage the opponent with strong direct strikes?

You should then fence from the guard that suits your fighting style the best. Obviously, the three guards introduced in the article  are hardly an exhaustive description of Fiore’s tactics, and other guards that are not discussed here can perform opportunistic actions such as binding the opponent’s sword or holding the center.

We can therefore confidently claim that Fiore’s guards first and foremost define an actual tactical framework. It may lack the comfort and canned answers that other more prescriptive systems provide, but I would argue it’s one of its main qualities: acknowledging that fencing is an open problem that can’t be solved with a one-size-fits-all solution.

Learning to take the blame

Most HEMA groups are led by good-willed volunteers who often try their best to make learning the art of historical fencing an enjoyable endeavor. But teaching HEMA can be a thankless job, because it lacks a well-established pedagogy. I will discuss in this article my own experience dealing with issues commonly encountered by budding HEMA instructors.

A teacher’s compendium

Whenever I introduce a new technique, my students want to come up with a counter.

Fencing is not a game of rock-paper-scissors. But it may be perceived as such if a curriculum over-emphasizes complex choreographed sequences of techniques. It fortunately only takes a few well-designed competitive drills to realize that knowing a technical motion is one thing, and being able to set it up is another.

The most clever parry will fail to stop a strike performed at the right measure, at the right moment. Make sure that your students understand a technique has to be set up and contextualized, and they’ll be too busy trying to make it work against a non-compliant opponent to come up with absurd counters. They may instead come up with practical solutions that will pave the way for latter lessons.

My students want to win the drill they’re practicing.

Not being able to train a technique because your partner is somehow always preemptively countering it instead of feeding you the proper stimulus can be a frustrating experience. But is a drill that actively trains one of the two fencers to fail a worthwhile exercise? Or would it instead hinder one’s progress?

Consider any choreographed sequence featuring an attack that is countered: no matter how much you emphasize proper form, any offensive action that leaves the attacker in a worse situation than they were before striking is suicidal. A cautious fencer would instead try to use preparatory actions to bait the opponent out of guard, or creep close enough that the strike can no longer be countered. Thus, can you blame your students for unconsciously trying to avoid a mistake?

There is very little value in learning how to perform or deal with straw man attacks that were never meant to hit in the first place – not to be mistaken with genuine attacks performed with improper form. Remember that a technical drill is not a fight, and it needs not look like one to be valuable. Keep it simple and focused on a few simple motions by abstracting everything that isn’t directly relevant to the action at hand, and your students will be far less likely to mistake the exercise for an actual fencing situation. As an example, you may remove footwork entirely if your current focus is first and foremost on bladework.

My students keep doubling each other and refuse to fight cleanly.

Yet your students may actually be trying their best with the tools you gave them. Clean fighting is a consequence of a deliberate, purposeful approach to fencing: each action is properly set up, the opponent’s intentions are analyzed and taken into account, and each motion is part of a bigger plan. A good fencer understands why a given action succeeds or fails.

This approach, however, can’t be improvised and has to be taught beforehand. A tactical and psychological framework must therefore be part of your curriculum and trained regularly, allowing your students to hone their decision-making, analysis, and planning skills. Should it be found wanting, then your sparring sessions may end up looking like two blind people wildly flailing at each other, and administering push-ups as a form of punishment will not change that.

My students form is great when they drill, but it breaks down when they spar and they can no longer apply any technique.

Form does not exist in a vacuum. It is meant to support a functional motion with a definite goal. But if the motions you show do not bring any tangible benefit to a fencer, then they will be discarded in favour of more natural ones. Assuming you are reasonably confident about the soundness of your teachings, then the issue may lie in your students inability to understand when a technique is relevant or not.

As an instructor, you should easily be able to demonstrate why a given posture or play is useful by contextualizing it: set up a simple exercise where proper application of the canonical movements immediately improves one’s success rate. As an example, playing games with a heavy emphasis on footwork may help the students understand the benefits of using a stable stance and small, fast steps.

My interpretation is fine, but we’re not good enough yet to apply it.

How can you know an interpretation is fine if you can’t test it? A statement that isn’t falsifiable should be considered with the utmost suspicion. Keep in mind HEMA is an experimental art: no matter how great you imagine your interpretation to be, your intuition may be wrong. Thus, if you can’t prove a given technique is sound, you may want instead to work on its prerequisites until you master them.

The hallmark of a good interpretation (assuming it fits the source and its historical context) is that it helps a fencer grow instead of needlessly constraining them. The problem-solving approach is therefore a good way to keep your fencing sane: if a given play does not help your students as much as you expected, then you should probably question your own understanding of it rather than double down on your current interpretation.

The impostor syndrome

By now, it should be rather obvious that I have been playing devil’s advocate from the very beginning of this article. Problematic students do exist: they may ignore your instructions, show off dubious skills learned on Youtube at your expense, or pose a physical and moral threat to other fencers. But we should always be ready to question our own interpretations and be aware of our limited knowledge, rather than blame good-willed students for failing to live up to our unrealistic expectations.

One way or another, most HEMA instructors are impostors, and I’m no exception to this rule. It took Fiore forty years to learn a living art under various experienced masters; a few self-taught enthusiasts are not going to spontaneously resurrect it using nothing but old fencing books.

[Getty 1v] I, Fiore, know how to read, write, and draw, and have books on our subject, a subject I have studied for at least forty years. Yet, I do not consider myself to be a perfect Master of this art, although some of the great lords who have been my students do hold me in such regard. (tr. by Tom Leoni and Greg Mele)